Integrating Artificial Intelligence and Data Science for Financial Decision-making and Business Performance Optimization
Patricia Ugochi Uzoma
Department of Business Management Administration, Clark University, Worcester, United States.
Kehinde Akinwale
Department Business Analysis, New Jersey City University, Jersey City, United States.
Blessing Itodo
Department of Information Science, University of Arkansas at Little Rock, Little Rock, United States.
Gideon Mawuli Gozah
College of Business and Analytics, Southern Illinois University, Carbondale, USA.
Ahmed Idowu Agbelejoye
Department of Mathematics and Statistics, Mississippi State University, Mississippi State, United States.
Confidence Adimchi Chinonyerem *
Department of Accountancy, Abia State Polytechnic, Aba, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Artificial Intelligence (AI) and Data Science are increasingly used to support corporate financial decision-making, yet uncertainty remains regarding how analytical maturity and AI adoption translate into measurable business performance. This study developed and evaluated an integrated decision-support framework using a synthetic dataset of 1,200 enterprise profiles representing six industrial sectors over the period 2019–2025. The analytical design combined Multiple Linear Regression, Support Vector Regression, Random Forest, Gradient Boosting Machine, Extreme Gradient Boosting (XGBoost), Artificial Neural Networks, Structural Equation Modelling, and SHapley Additive exPlanations (SHAP). Model performance was assessed through train-test partitioning, hyperparameter optimisation, cross-validation, robustness testing, sector-specific validation, and sensitivity analysis. XGBoost achieved the strongest predictive performance, explaining 91.4% of the variance in the Business Performance Index (R² = 0.914), with an RMSE of 3.623 and an MAE of 2.541. Structural analysis indicated that the Financial Decision Score mediated the relationship between AI adoption and business performance, with an indirect pathway coefficient of β = 0.218. SHAP analysis identified a nonlinear threshold pattern, with stronger predicted performance effects when the AI Adoption Index exceeded approximately 60 points. The framework remained comparatively stable under noise injection and across industrial sectors. Overall, the findings indicate that business value is associated not simply with AI adoption, but with its systematic integration into financial decision-making processes.
Keywords: Artificial intelligence, corporate financial decision-making, data science, decision support systems, explainable artificial intelligence, machine learning, structural equation modelling